The greenhouse gas value of ecosystems

Global Change Biology (2010), doi: 10.1111/j.1365-2486.2010.02220.x The greenhouse gas value of ecosystems K R I S T I N A J . A N D E R S O N - T E ...
Author: Sharyl Reynolds
8 downloads 0 Views 493KB Size
Global Change Biology (2010), doi: 10.1111/j.1365-2486.2010.02220.x

The greenhouse gas value of ecosystems K R I S T I N A J . A N D E R S O N - T E I X E I R A *w and E VA N H . D E L U C I A *w z *Institute of Genomic Biology, University of Illinois at Urbana-Champaign, IL 61801, USA, wEnergy Biosciences Institute, University of Illinois at Urbana-Champaign, IL 61801, USA, zDepartment of Plant Biology, University of Illinois at Urbana-Champaign, IL 61801, USA

Abstract As society faces the urgent need to mitigate climate change, it is critical to understand how various ecosystems contribute to the climate, and to express these contributions in terms that are meaningful to policymakers, economists, land managers, and other nonscience interest holders. Efforts to mitigate climate change call for quantification of the full greenhouse gas (GHG) effects of land use decisions, yet we lack an appropriate metric of the full GHG implications of maintaining a given ecosystem over a multiple year time frame. Here, we propose the concept of greenhouse gas value (GHGV) of ecosystems, which accounts for potential GHG release upon clearing of stored organic matter, annual GHG flux, and probable GHG exchanges resulting from disturbance. It treats these ecosystem– atmosphere exchanges in a time-sensitive manner, thereby providing an appropriate framework for computing of the GHG consequences of any land use decision. To illustrate this concept, we provide estimates of the GHGV of various biome types (based on data compiled from the literature), disturbance regimes, and decisions on the treatment of time. We show that natural ecosystems generally have high GHGV’s, whereas managed ecosystems generally have lower or negative GHGV’s; that GHGV decreases with increasing probability of disturbance, and that decisions on the treatment of time can be important, affecting some ecosystem types more strongly than others. In addition, we show how GHGV may be used to quantify the full GHG effects of land-use or land-cover change in a thorough and rigorous manner. Finally, we provide comparisons of GHGV to other major paradigms for valuing the GHG contributions of ecosystems, showing that – for many purposes – GHGV is the most appropriate method of quantifying the GHG services of ecosystems. Keywords: biofuels, biomes, carbon dioxide (CO2), disturbance, ecosystem services, ecosystem-atmosphere exchange, fire, landuse/land-cover change, methane (CH4), nitrous oxide (N2O)

Received 25 January 2010 and accepted 8 March 2010

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

Introduction As society faces the urgent need to mitigate climate change, it is increasingly important to understand the contributions of terrestrial ecosystems to climate. This is particularly important in the face of rapid land-use change, which may accelerate because of intensifying pressure from agriculture (e.g., Searchinger et al., 2008). Substantial changes in ecosystem structure may also occur as a result of climate change, at times causing complete displacement of biomes (e.g., Scholze et al., 2006; Anderson-Teixeira et al., in review). Such land cover changes typically entail considerable greenhouse gas (GHG) release upon clearing of the original ecosystem (e.g., Fearnside, 2000; Fargione et al., 2008) or longterm changes in annual GHG flux (e.g., Robertson et al., 2000). Land management decisions thereby contribute Correspondence: Evan H. DeLucia, Institute of Genomic Biology, University of Illinois at Urbana-Champaign, IL 61801, USA. tel. 1 1 217 333 0860, fax 1 1 217 244 7246, e-mail: [email protected]

meaningfully to global GHG budgets (Houghton, 2007; Smith et al., 2007), and efforts to mitigate GHG emissions from terrestrial ecosystems therefore require rigorous accounting for the GHG services of ecosystems. There are four main ways in which the GHG services of ecosystems are commonly valued (Table 1). First, ecosystems may be valued for their storage of organic matter that would be released as GHG’s upon land clearing (Dixon et al., 1994; Miles & Kapos, 2008). This value is the focus of the UN Framework Convention on Climate Change’s (UNFCCC) program on Reducing Emissions from Deforestation and Forest Degradation (REDD-plus; Miles & Kapos, 2008; UNFCCC, 2008, 2009), which would provide a financial incentive for developing countries to reduce deforestation. Second, ecosystems may be valued based on measured or projected GHG flux (e.g., CO2 sequestration or N2O release) over a time span of interest (e.g., Robertson et al., 2000; Lal, 2004; Righelato & Spracklen, 2007; Anderson-Teixeira et al., 2009; Smeets et al., 2009). In some cases, when a multi-year time frame is 1

r 2010 Blackwell Publishing Ltd

2220

G C B

2 2 2 0

Journal Name

Manuscript No.

B

Dispatch: 12.4.10 Author Received:

Journal: GCB No. of pages: 14

CE: Mamatha/Tanuja PE: Dherian/Gk

GCB

2220 Quantification of the GHG’s (often CO2 only) that would be released if the ecosystem were to be cleared. Many applications consider CO2 only, and do not necessarily quantify all organic matter pools Annual ecosystem-atmosphere GHG exchange, sometimes extrapolated over multiple years Annual reporting of GHG emissions – generally from managed ecosystems (e.g., IPCC guidelines for reporting under UNFCCC) Combined contributions from storage and flux over a multiyear time span without proper accounting for the treatment of time. Applications (e.g., biofuels life cycle analyses) vary in comprehensiveness Accounting for flux, storage, and probable effects of disturbance over a multiyear time period with proper accounting for time

Organic Matter Storage

Annual GHG Inventory

U.S. Renewable Fuels Standard (US EPA, 2009); California Low Carbon Fuel Standard (California EPA, 2009)

1

1

w

1

1

na



1

1*





1

/ 1

/ 1

Original version of REDD (Miles & Kapos, 2008; UNFCCC, 2008)

Chicago Climate Exchange (CCX, 2009); Renewable Transport Fuel Obligation (UK Parliament, 2007) IPCC Guidelines for National Greenhouse Gas Inventories (IPCC, 2006); REDD-plus (UNFCCC, 2009)



Multiyear

Policy examples

Organic matter storage

1

1

1

1



Annual GHG flux

1



1*





Probable disturbance effects

storage and risk of disturbance per se do not affect these inventories. wU.S. EPA’s Renewable Fuel Standard proposed rule (US EPA, 2009) applied discounting for their 100-year analyses, but was otherwise insensitive to the timing of emissions.

*Annual GHG inventories count storage and disturbance in the sense that emissions from GHG emissions from clearing or fire would be counted in annual inventories; however,

GHGV

Storage and Flux

GHG Flux

Description

Metric

Sensitive to timing of emissions

Table 1 Comparison of metrics of the GHG contributions of ecosystems

Accounting for

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

2 K. J. ANDERSON-TEIXEIRA & E. H. DELUCIA

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

G R E E N H O U S E G A S VA L U E O F E C O S Y S T E M S

GCB

considered, GHG fluxes are weighed by the year in which they occur. A prominent application of this metric is the Chicago Carbon Exchange, where land management-induced mitigation of annual GHG emissions can be sold as carbon offsets (CCX, 2009). Similar methods of measuring domestic carbon offsets may be implemented under a federal cap-and-trade policy in the United States [e.g., American Clean Energy and Security Act (H.R. 2454); US House of Representatives, 2009]. This method of valuing ecosystems also was used in early biofuel life-cycle analyses that did not consider loss of carbon storage through land use change (e.g., Tilman et al., 2006; Adler et al., 2007; UK Parliament, 2007). Third, contributions from both release of stored organic material and annual GHG flux may be counted through inventories of ecosystem-atmosphere exchanges over relatively short (e.g., annual) time periods. This is the approach used for national inventories of GHG emissions from managed land as described by the Intergovernmental Panel on Climate Change (IPCC, 2006). Fourth, the combined contributions of storage and flux may be evaluated over longer time periods, as is commonly done in life cycle analyses that seek to determine the net GHG effects of biofuel production (e.g., Searchinger et al., 2008; California EPA, 2009; Melillo et al., 2009; US EPA, 2009). All four of these approaches for estimating the GHG value of ecosystems face certain limitations (Table 1). Those that consider only organic matter storage or only flux – while appropriate for certain applications – miss the value associated with the other. For example, basing a forest’s value on the GHG release that would be incurred by its destruction misses the value associated with annual C sequestration, whereas basing an ecosystem’s value solely on its GHG flux misses the value of its stored organic material. These approaches typically do not employ an appropriate strategy for quantifying the effects of emissions that occur over multiple years (e.g., annual fluxes, decomposition of wood after forest clearing). Multiyear paradigms often err in their treatment of the timing of emissions, as proper accounting for time over multiyear time frames precludes a simple summation of emissions expressed as CO2equivalents (O’Hare et al., 2009). Annual inventories of GHG exchange are valid for purposes such as reporting of national GHG emissions under the UN Framework Convention on Climate Change (IPCC, 2006); however, they deal with accomplished emissions on a year-toyear basis and do not provide the structure for understanding the full climate impacts of land-use/landcover change (henceforth, LULCC) over a multiyear time frame. Finally, these approaches typically do not account for probable effects of natural disturbance (but see IPCC, 2006), which may strongly affect the value of

preserving an ecosystem. For example, it may be less beneficial to preserve a forest that is likely to burn, as a burn would result in significant GHG release. Thus, there is a strong need for a method of valuing the GHG contributions of ecosystems that accounts for contributions from organic matter storage, annual flux, and probability of disturbance – and does so over multiple years in a time-appropriate manner. Here, we propose a quantitative metric for the greenhouse gas value of ecosystems (GHGV), which quantifies the value of maintaining an ecosystem over a multiple-year time frame. It incorporates changes in ecosystem-atmosphere GHG exchanges that would occur if the ecosystem were to be cleared, including release of GHG’s through oxidation of stored organic material, displaced annual GHG flux, and displaced probable effects of disturbance. Through proper accounting for the timing of emissions, it quantifies the total radiative forcing that would occur upon clearing of the ecosystem. GHGV is expressed relative to the effects of a pulse CO2 emission such that its units are expressed in CO2-equivalents, which are useful in that they serve as a currency for emissions regulation and offset trading. Following the theoretical development of GHGV, we use it to provide illustrative calculations of GHGV for various biomes, disturbance regimes, and decisions on the treatment of time. We also discuss how the GHGV’s of two ecosystems may be combined to quantify the full GHG effects of LULCC.

Theoretical development of GHGV We define an ecosystem’s GHGV as the total benefit of avoiding radiative forcing from GHG’s through maintenance of 1 ha of the ecosystem. GHGV incorporates potential GHG release upon clearing of stored organic matter, the annual flux of GHG’s from the ecosystem to the atmosphere, and probable GHG exchange resulting from disturbance. Ecosystem–atmosphere GHG exchanges are counted over an emissions time frame tE (years), which is an arbitrarily chosen time span over which the ecosystem’s GHG impact is assessed (see below for discussion of tE). Because GHG’s remain in the atmosphere – and thereby impact the climate – for many years following their release, it is generally desirable to evaluate the climate impact of ecosystem– atmosphere GHG exchange (i.e., cumulative radiative forcing) over a longer analytical time frame, tA (years; analogous to ‘time horizon’ for global warming potentials; Forster et al., 2007). Radiative forcing may be weighted over time according to its potential to trigger dangerous climate change; here, we employ a weighting function, w(tA) (default w 5 1), which is discussed below. Analogous to the global warming potential

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

3

4 K. J. ANDERSON-TEIXEIRA & E. H. DELUCIA metric (Forster et al., 2007) but differing in its treatment of a multiple-year time emissions time span, GHGV is expressed in CO2-equivalents (Mg CO2-eq ha1) through comparison with the cumulative radiative forcing that would arise from a pulse emission of CO2 at time zero. Formally, GHGV is as follows: R tA 

GHGVttAE

¼

RFðtA Þ ¼

X

air in the atmosphere (A 5 1.78  108 billion kmol). The portion of remaining in the atmosphere at time tA, rx, is calculated based on the decay of a pulse of GHG x over time (see Forster et al. (2007) for rx’s used here). Ix is as follows:

Ix ðtE Þ ¼ Sx ðtE Þ  Fx ðtE Þ  Dx ðtE Þ:



RFtE ðtA ÞwðtA Þ dtA  R ttAA¼o  GHG tA ¼0 RFpCO2 ðtA ÞwðtA Þ dtA

:

ax CtxE ðtA Þ:

ð4Þ

ð1Þ

E Here, RFtGHG ðtA Þ (nW m2 ha1 ecosystem yr1) is the additional radiative forcing at time tA that would arise from GHG’s released over time span tE following the clearing 1 ha of the ecosystem, and RFpCO2 ðtA Þ (nW m2 yr1) is the additional radiative forcing that would arise from a pulse emission of 1 Mg CO2 at tA 5 0. The numerator of this equation gives the ecosystem’s value in terms of radiative forcing (nW m2 ha1 ecosystem), while the denominator translates it into CO2-equivalents. For each year in the analytical time span, additional E and RFpCO2 ) is calculated by radiative forcing (RFtGHG summing the radiative forcings from of all pertinent GHG’s (subscripted x):

Here, Sx, Fx, and Dx (kmol x ha1 yr1) are the potential release of GHG’s from stored organic material upon land clearing, displaced annual flux, and displaced probable emissions from natural disturbance, respectively. These are detailed below.

ð2Þ

x

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

Here, ax is the effective [For methane, this is calculated by multiplying the radiative efficiency of methane (3.7  105 nW m2 ppb1; Forster et al. 2007) by 4/3 to correct for indirect effects from enhancements to ozone and stratospheric water vapor. This gives very results to the IPCC’s global warming potential estimates for 20-, 100-, and 500-year time horizons (Forster et al. 2007)] radiative efficiency of GHG species x (aCO2 5 1.4  104, aCH4 ¼ 4:9  105 , and aN2 O ¼ 3:03  106 nWm2 ppb1 ; Forster et al., 2007). This is multiplied by the additional atmospheric abundance of GHG x attributable to clearing of 1 ha of the ecosystem, CtxE (ppb ha1 ecosystem yr1), which varies as a function of time. In turn, CtxE ðtA Þ is determined based on the change in GHG fluxes that would occur as a result of clearing 1 ha of land, integrated over the emissions time period of interest (tE):

CtxE ðtA Þ ¼

Z

minðtE ;tA Þ tE ¼0

  Ix ðtE Þ rx ðtA  tE Þ dtE : A

ð3Þ

Here, Ix (kmol x ha1 yr1) is the input of GHG x from the ecosystem to the atmosphere and A is the moles of

Fig. 1 Example of how a time series of ecosystem-atmosphere GHG exchange translates into GHGV for a tropical forest cleared by burning (see SI; Fig. 2). This illustration uses a 50-year emissions time span (tE), no discounting [Eqn (1)], and no probability of disturbance [Eqns (4) and (6)]. (a) 50 year time series of ecosystem-atmosphere exchange of for CO2, CH4, and N2O [Ix; Eqns (3) and (4)]; (b) resulting additional atmospheric abundance of these GHG’s through time [CtxE ; Eqns (2) and (3)]; (c) resulting radiative forcing from each GHG through time [Eqn (2)]; (d) cumulative radiative forcing from all GHG’s combined E [RFtGHG ; Eqns (1) and (2)]; (e) GHGV [Eqn (1)] as a function of the analytical time frame. Also illustrated are the effects of a 1 Mg CO2 pulse at time 0, which is used as a basis of comparison to express GHGV in CO2-equivalents [Eqn (1)]. For display purposes, some values are multiplied by 100 or 1000 (see legends).

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

G R E E N H O U S E G A S VA L U E O F E C O S Y S T E M S Figure 1 illustrates how the time series of ecosystematmosphere GHG exchange, Ix(tE), translates into GHGV based on Eqns (1)–(4).

Contributions of stored organic matter (S)

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

Ecosystems store organic material that is oxidized upon land clearing – either immediately through combustion or over time through decomposition – resulting in GHG release. We define the storage of ‘vulnerable’ organic material as all aboveground and belowground biomass, dead wood, the organic layer consisting of litter or peat, and any soil organic material (SOM) in the top meter of mineral soil that would be released through long-term annual tillage. Such a SOM baseline is necessary because a large portion of SOM exists in relatively stable forms or deep soil layers and would be minimally affected by LULCC. Annually tilled croplands provide an appropriate baseline, as their soils are the most carbon-depleted among common land use types (Guo & Gifford, 2002), containing on average approximately 30% less SOM in the top meter than their undisturbed counterparts (Davidson & Ackerman, 1993; Guo & Gifford, 2002; Murty et al., 2002). Thus, in cases where annually tilled ecosystems are not available for comparison, a reasonable default for native ecosystems is to include 30% of SOM in the top meter of mineral soil. Note that organic matter storage is not necessarily equal to the reduction in organic material that occurs during LULCC; while the baseline serves to allow consistent comparison across sites, LULCC calculations will consider the change in organic matter storage (see below for discussion of land-use/land-cover change). Translation of organic matter storage into GHG release as a function of time [Sx(tE); kmol x ha1 yr1; Eqn (4)] is sensitive to the mechanism of land clearing and the distribution of organic material among different pools (e.g., combustible biomass, roots, soil organic matter). For each pool (subscripted p), GHG release is separated into an initial (i.e., tE 5 0) release through combustion – if fire is used for land clearing – and a subsequent (tE40) release of the remaining material through decomposition – either onsite or off-site as wood products:

Sx ðtE Þ ¼

X p

( OMp

! fpc Ecx;p tE ¼ 0 : ð5Þ ð1  fpc ÞEdx;p dp ðtE Þ tE > 0

Here, OMp (Mg dry matter ha1) is the organic matter in each pool, and fcp and (1fcp) are the fractions of organic matter in that pool oxidized through combustion and decomposition, respectively. The terms Ecx;p

and Edx;p (kmol x Mg1 dry matter) represent the proportional release of GHG x for each unit of biomass oxidized through combustion or decomposition, respectively, and dp(tE) is the proportion of organic matter in pool p that decomposes each year (0  dp(tE)  1). Typically, dp(tE) will be characterized by an exponential decay function. We note that – when fire is used in the process of land clearing – S will be particularly sensitive to burn characteristics (fcp and Ecx;p ), which may be highly variable within a single ecosystem type. Biomass burning releases a wide range of trace GHG’s, including indirect GHG’s such as carbon monoxide (CO), non-methane volatile organic compounds, and mono-nitrogen oxides (NOx) (e.g., Andreae & Merlet, 2001). In theory, the climate impacts of all GHG’s should be assessed; however, this is complicated by high uncertainty in both emission factors and global warming potential estimates for some of the less abundant GHG’s (e.g., Andreae & Merlet, 2001; Forster et al., 2007). For release through decomposition, the most important GHG species is CO2, although decomposition of stored organic material also sometimes releases significant quantities of CH4 (i.e., from termite-mediated decomposition) or N2O (i.e., peatland organic soils).

Contributions of annual flux (F) The GHGV of ecosystems also includes their potential to sequester or release GHG (e.g., Robertson et al., 2000; Robertson & Grace, 2004). This flux of GHG’s to or from the ecosystem (F; Eqn (4); positive value indicates release to atmosphere) is dominated by the GHG’s CO2, CH4, and N2O. The net flux of CO2, or net ecosystem exchange, is dominated by carbon release through respiration and uptake through photosynthesis, but also includes non-respiratory fluxes (e.g., fire, ultraviolet oxidation of organic matter; Chapin et al., 2006). It generally approximates – but is not identical to – the net change in ecosystem carbon storage (Chapin et al., 2006). Any carbon removed from the ecosystem through harvest that will be returned to the atmosphere within a short time frame (e.g., through human consumption and respiration, or through biofuels burning) should be subtracted from FCO2 . Fluxes of CH4 and N2O are typically small in quantity but potent because of the high global warming potential of these gases, and thereby can strongly affect the GHGV of some ecosystems. For managed ecosystems, a complete GHG budget must also include associated anthropogenic emissions, such as those from farm machinery (i.e., fuel combustion emissions) and those associated with agricultural inputs (e.g., emissions from lime and fertilizer production; Robertson et al., 2000; West & Marland, 2002; Robertson & Grace, 2004).

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

5

6 K. J. ANDERSON-TEIXEIRA & E. H. DELUCIA Calculating GHGV requires values for Fx over the entire time span of interest (Eqn (4)). As a complete time course of GHG fluxes is methodologically difficult to obtain, it will often be necessary to approximate F as a constant through time, or, for aggrading ecosystems, as simple step function (e.g., see Supporting Information). In reality, however, F changes continuously as ecosystems age (Odum, 1969; Robertson et al., 2000; Law et al., 2003; Litvak et al., 2003; Bond-Lamberty et al., 2004; Luyssaert et al., 2008). We note that the course of GHG fluxes as a function of ecosystem age remains controversial for CO2 (e.g., Bond-Lamberty et al., 2004; Magnani et al., 2007; Luyssaert et al., 2008) and has not yet been extensively studied for CH4 and N2O (but see Prieme´ et al., 1997; Robertson et al., 2000; Peichl et al., 2009). In some cases, resolving these dynamics will be key to producing reliable estimates of GHGV.

Probable contributions from disturbance (D)

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

Natural disturbances are stochastic, yet stand to strongly impact the GHGV of ecosystems. The possibility of multiple types of disturbances of varying severity and frequency, as well as repeat disturbances during the time span of interest makes calculation of the full effects of disturbance complex. In theory, relatively frequent, small natural disturbances (e.g., small-scale mortality, tree-fall events, surface fires) are incorporated into annual GHG flux measurements. The extent to which this actually occurs is mixed. Eddy-covariance flux measurements and biometric estimates of annual carbon flux capture individual mortality, tree-fall events, and – to the extent that researchers do not manipulate their frequency – surface fires. Chamber measurements (e.g., for trace GHG’s) are more likely biased to avoid these disturbances. Here, we incorporate disturbance by adjusting for the probabilistic impacts of large, infrequent disturbances (e.g., stand-clearing forest fires, hurricanes, insect outbreaks). Our approximation assumes that at most one disturbance occurs during the time span of interest, which is reasonable when the disturbance rate (rD; yr1) is low and/or the time span of interest (tE) is short (i.e., rD tE  1), or when repeat disturbance is unlikely to occur shortly following a disturbance. It also assumes that disturbance rate remains constant over the time span of interest. The expected contribution of GHG x from disturbance, Dx, is as follows: D Dx ðtE Þ ¼ rD ½SD x ðtE Þ  Fx ðtE Þ

ð6Þ

Here, SxD (kmol x ha1 yr1) is the expected GHG release from disturbance, and FxD(tE) (kmol x ha1 yr1)

is the probable net change in cumulative GHG flux arising from recent disturbance. For each year within the emissions time span, potential GHG release through disturbance, SxD, is calculated by summing the possible GHG exchange from disturbance in that or any previous year [this is then adjusted for the probability of disturbance in Eqn (6)]: SD x ðtE Þ ¼

X p

0 @

tE X t ¼1

0 @OMD p

8 < fpcD Ecx;p h i  : ð1  fpcD ÞEdx;p dD p ðtE  t Þ

t ¼ tE t < tE

11 AA

ð7Þ 1 Here, OMD p (Mg ha ) is the storage of organic matter in pool p subject to release by the disturbance. It will often be less than the organic matter subject to release through anthropogenic clearing [OMp; Eqn (5)], as mortality may not be complete and the soil will generally remain relatively undisturbed. Likewise, the portion of biomass that burns (fpcD ) and the decay function of remaining biomass (dD p ) may differ from those that describe land-clearing [Eqn (5)]. The variable t refers to the year in which potential disturbance occurs. During the period of recovery following a disturbance, ecosystems have altered GHG flux rates (e.g., Bormann & Likens, 1994) and this effect is quantified by FxD(tE):

R R FD x ðtE Þ ¼ minðtE ; t ÞðFx  Fx Þ

ð8Þ

Here, tR is the time it takes for the ecosystem to ‘recover’ to its pre-disturbance state, and the term min(tE, tR) reflects the fact that any disturbance since the beginning of the time span and within this recovery period would impact the flux at time tE. The term (FxRFx) expresses the difference in annual flux between the recovering ecosystem and a counterpart that has not been recently disturbed. This assumes a constant difference between FD and F over time, which is not realistic, but is a useful first approximation. Disturbance has an important impact on GHGV, and the probability of disturbance should not be ignored in land use decisions. At the same time, challenges remain to accurately quantifying the effects of disturbance. First, disturbance rate is not necessarily easy to predict, as climate change and other anthropogenic impacts are altering disturbance frequencies (e.g., Westerling et al., 2006). Additionally, our current treatment of disturbance accounts for only one disturbance type and size, and does not allow repeat disturbances in the time frame of interest. While valuable in quantifying the effects of infrequent stand-clearing disturbances, our model [Eqns (6)–(8)] will require further development

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

G R E E N H O U S E G A S VA L U E O F E C O S Y S T E M S to capture the impacts of a full suite of possible disturbances.

There are three time-related issues that can impact the value of GHGV: the choice of an emissions time frame [tE; Eqn (3)], the choice of an analytical time frame [tA; Eqn (1)], and the (optional) application of a weighting function, [w(tA); Eqn (1)].

fluxes over tE, it should exceed tE by at least 50 years. Many applications select tA 5 100 years through use of the IPCC’s global warming potential values for a 100year time horizon (Forster et al., 2007). Thus, for most applications, use of tA 5 100 years would be a logical choice that is consistent with other literature. If the climate change impacts of land use decisions are being evaluated relative to a certain target date, tE and tA should both equal the number of years until that target date.

Emissions time frame (tE)

Weighting by time of emission [w(tA)]

GHGV is dependent upon the time frame over which ecosystem-atmosphere GHG exchange is counted (emissions time span;tE). In an immediate sense, the GHGV’s of ecosystems are strongly dependent on GHG release from stored organic material [S; Eqn (5)] – particularly combustion emissions – and, as tE increases, contributions from annual flux (F) and probable effects of disturbance (D) become increasingly important. There is no ‘correct’ tE; rather, the appropriate time frame depends upon the nature of the application. There are, however, a few criteria that should be observed in the selection of tE. First, tE should not be too short to capture the decomposition fated to occur by the clearing of the ecosystem. The tE required to capture this will depend upon organic matter storage and decomposition rates [Eqn (5)]. On the other hand, tE could be excessively long if it exceeds the time frame over which conditions can be expected to remain reasonably predictable. In particular, the high likelihood that global change will impact annual GHG exchange (F) and disturbance frequency in many ecosystems (e.g., Westerling et al., 2006; Field et al., 2007) implies increasing uncertainty at longer time scales. Finally, for LULCC applications, one reasonable criterion for tE would be the expected duration of the new ecosystem. For example, analyses of the GHG impacts of biofuels-related land use may select tE based upon the expected duration of ethanol production (e.g., Righelato & Spracklen, 2007; Searchinger et al., 2008; O’Hare et al., 2009).

It may be desirable to place greater weight on current than on future emissions. Earlier emissions, for example, may be more likely to determine the fate of climate change in that they could trigger feedback mechanisms or push the climate system past critical damage thresholds (e.g., Lenton et al., 2008). In addition, because society tends to place more value on near-term than on long-term costs and benefits, economic and policy applications often apply an annual discount rate to future emissions (e.g., Kim et al., 2008). Weighting by the timing of emissions can easily be applied to GHGV through w(tA) [Eqn (1); w 5 1 for no weighting]. One weighting function commonly used in economics is wðtA Þ ¼ ð1=ð1 þ rÞÞtA , where r is the annual discount rate (typically, ro0.1; e.g., Kim et al., 2008; US EPA, 2009). Criteria for selecting w(tA) are beyond the scope of this analysis.

Treatment of time

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

7

Analytical time frame (tA) GHGV is also dependent upon the time over which the climate impacts are evaluated (analytical time frame;tA). Whereas cumulative radiative forcing continues to increase over hundreds of years, normalization by a CO2 pulse causes GHGV of most ecosystems to stabilize within about 150 years after the cessation of emissions. As with tE, the choice of tA is subjective, but should be guided by several considerations. Logically, tA must be greater than or equal to tE. To capture the full effects of

Illustrative GHGV calculations To provide estimates of GHGV of various ecosystem types and to illustrate the effects of disturbance and the treatment of time (Figs 2–4), we compiled published estimates of pertinent variables [Eqns (5)–(8)] for a variety of ecosystem types (see Supporting Information for details and parameter estimates). GHGV was calculated according to the above equations using Matlab 7.7.0 (Mathworks Inc.). Matlab code and an accompanying spreadsheet (Microsofts Excel) are provided as supporting information. Estimates included GHG species CO2, CH4, and N2O, which are the most prominent GHG’s exchanged between ecosystems and the atmosphere.

GHGV estimates for various biomes We calculated GHGV for a variety of native, aggrading (recently disturbed or abandoned), and managed ecosystem types (Table S1). For consistency, all examples represent a disturbance of fire followed by on-site decomposition of unburned material. The probability

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

8 K. J. ANDERSON-TEIXEIRA & E. H. DELUCIA

Fig. 2 Contributions to GHGV from (a) storage of materials vulnerable to release as GHG’s upon land clearing [S; Eqn (5)] and (b) displaced flux of CO2, CH4, and N2O, and anthropogenic emissions (Fanth). These are combined to yield (c) greenhouse gas value for a 50-year emissions time span and 100 year analytical time span, GHGV50 100. Here, effects of disturbance are not included, and no discounting is applied. Values are presented in tabular form in Table S9.

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

Fig. 3 Influence of the probability of a stand-clearing forest fire on GHGV50 100 of tropical, temperate, and boreal forests. Storage and flux components are calculated as in Fig. 2. See text for details.

of natural disturbance was not incorporated into these estimates (but see below for examples with natural disturbance). In this illustration, we selected an emissions time span of 50 years and an analytical time span of 100 years, and discounting was not applied (see below for examples with variation in time variables). Contribution to GHGV from stored organic material [Eqn (5); Fig. 2a; Table S9] is closely linked to total organic matter storage (Table S2), and also is affected by burn characteristics (fcp and Ecx;p ; Eqn (5); Table S3) and the timing of release of decomposing organic material

[dp(tE); Eqn (5); Table S4]. The contribution to GHGV from S ranged from minimal values in current or abandoned cropland (o5 Mg CO2-eq ha1) up to 3310 Mg CO2-eq ha1 in tropical peat forests. This was dominated by potential CO2 release, which represented at least 90% of the contribution from S in all ecosystems. Potential releases of CH4 and N2O contributed up to 3% and 8%, respectively. Contributions to GHGV from annual GHG flux tended to be positive in natural ecosystems (with the exception of wetlands) and negative in managed ones, ranging from 524 Mg CO2-eq ha1 in aggrading tropical forests to 535 Mg CO2-eq ha1 in wetland rice (Fig. 2b; Table S9). It was not consistently dominated by any one GHG species. In unmanaged ecosystems with unsaturated soils, contributions from CO2 flux (Table S5) dominated and were consistently positive. That is, native ecosystems tended to be carbon sinks (albeit very high variability in NEE), as has been previously observed (Law et al., 2002; Luyssaert et al., 2007). Aggrading ecosystems were consistently carbon sinks, with C sequestration rates increasing from cool to warm climates (Fig. 2b; see also Anderson et al., 2006). In contrast, CO2 flux was minimal in managed ecosystems, where flux contributions from CH4 and N2O prevailed. Annual methane flux (Table S6) reduced GHGV50 100’s of all wetlands (up to 460 Mg CO2-eq ha1 in wetland rice), and also contributed meaningfully to the

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

G R E E N H O U S E G A S VA L U E O F E C O S Y S T E M S

Fig. 4 GHGV for five tropical ecosystem types as a function of (a) emissions time span (tE), (b) analytical time span (tA), and (c) annual discount rate, r, in the weighting function wðtA Þ ¼ ð1=ð1 þ rÞÞtA . For display purposes, values for tropical peat forest are multiplied by 0.1.

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

9

GCB

GHGV50 100’s of pastures. Its contributions were minimal for ecosystems with unsaturated soils and no livestock (3–5 Mg CO2-eq ha1). Because of high uncertainty associated with methane flux estimates from soil (Le Mer & Roger, 2001), uncertain CH4 contributions from plants (Houweling et al., 2006; Keppler et al., 2006), and high variability in cattle density, methane flux remains highly uncertain for many ecosystem types. Nitrous oxide flux (Table S7) contributed minimally to the GHGV of native and aggrading ecosystems (3 to 44 Mg CO2-eq ha1), but was important in managed ecosystems, where its contributions ranged from 36 Mg CO2-eq ha1 in tropical pastures to 82 Mg CO2-eq ha1 in croplands. In croplands, CO2 emissions associated with fertilizer production, lime, and fuel use by farm machinery (Table S8) contributed 25 to 63 Mg CO2eq ha1, depending strongly on assumed crop management practices. GHGV50 100 (Fig. 2c, Table S9) was consistently positive in natural ecosystems, with relatively modest values in dry or cold climates (o600 Mg CO2-eq ha1) and high values in forests (4950 Mg CO2-eq ha1 in tropical and temperate forests). Despite large methane releases from peatlands (outweighing the benefits of CO2 sequestration in northern peatlands; see also Roulet, 2000; Laine et al., 2006), they had the highest GHGV50 100’s (41600 Mg CO2-eq ha1) because of substantial storage of carbon in peat. Aggrading ecosystems had modest GHGV50 100’s, ranging up to 529 Mg CO2-eq ha1 in aggrading tropical forests. Despite negative contributions from FGHG, moderately grazed pastures had positive GHGV50 100’s owing to relatively high SGHG’s (Fig. 2). Crop ecosystems, on the other hand, had negative GHGV50 100’s (o80 Mg CO2-eq ha1 for both temperate and tropical crops), which were dominated by contributions from N2O and associated anthropogenic emissions. Thus, in general, unmanaged ecosystems and moderately stocked pastures are acting to mitigate GHG’s through storage of organic material and GHG uptake, whereas crop ecosystems are GHG sources.

We caution that the values presented here should not be used as off-the-shelf estimates of GHGV for any particular ecosystem. Influential variables such as organic matter storage, burn characteristics, cattle density, and crop management practices can vary by orders of magnitude within some of the biomes considered here. Therefore – particularly for influential and often easily measured variables such as these – GHGV should be calculated based on the best available data for the specific ecosystem(s) of interest.

Effect of forest fires on GHGV To illustrate the effects of natural disturbance on GHGV, we calculated the GHGV of tropical, temperate, and boreal forests across a range of stand-clearing fire frequencies (Fig. 3). Specifically, we considered fire frequencies [rD; Eqn (6)] ranging from 0 (no fires) to 0.015 yr1 (66 year fire cycle), which encompasses the range of fire frequencies typically observed for all three forest types (e.g., Turner & Romme, 1994; van der Werf et al., 2008). We assumed that any unburned biomass or woody debris would be committed to release through decomposition (all vegetation killed) but that SOC would be unaltered. For all three GHG species, we assumed annual GHG flux during recovery (FxR; Eqn (8)) to be equal to GHG flux of the corresponding aggrading ecosystems during the time of recovery [Eqn (8); tR estimated at 23, 72, and 75 years for tropical, temperate, and boreal forests, respectively; see Supporting Information for details]. As in the above example, we used values of tE 5 50 and tA 5 100, and no discounting was applied. Assuming no repeat disturbances, GHGV50 100 decreases linearly with fire frequency (Fig. 3). The main effect of disturbance arises from the probability that GHG’s will be released through combustion and subsequent decomposition [SD; Eqns (6) and (7)]. This is partially offset by increased C uptake in the recovering forest [FD; Eqns (6) and (8)]; however, even the most

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

10 K . J . A N D E R S O N - T E I X E I R A & E . H . D E L U C I A rapid recovery would not compensate for GHG releases through fire. Differences in the slope of the GHGV-fire frequency relationship (Fig. 3) are attributable to ecosystem characteristics. Specifically, probable release from stored organic matter [Eqn (7)] depends upon biomass storage and is larger for high-biomass temperate forests than for tropical or boreal forests (Fig. 2a). Additionally, contributions from disturbance depend strongly upon the difference between rates of carbon uptake in native and re-growing forests. Because forests re-grow most rapidly in the tropics (Anderson et al., 2006), there is a large increase in carbon uptake by recovering tropical forests (FD), thereby lessening the negative impacts of disturbance. It is for these reasons that GHGV50 100 decreases more rapidly with increasing fire frequency in temperate forests than in tropical or boreal forests.

Effects of time

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

To illustrate the effects of time on GHGV, we consider how GHGV changes with the emissions time span (tE), the analytical time span (tA), and the annual discount rate (r) for several ecosystem types that represent a variety of storage-flux combinations (Fig. 4). Specifically, we consider a tropical peat forest (extremely high organic matter storage released over a long time, large CO2 sink, large CH4 source), a tropical forest (high organic matter storage, moderate annual flux), an aggrading tropical forest (minimal organic matter storage, strong CO2 sink), a tropical cropland (minimal organic matter storage, large N2O source), and wetland rice (some organic matter storage, extremely high FCH4 ). Choice of tE has the strongest effect on ecosystems with high GHG flux or slow release of stored organic matter (Fig. 4a). The GHGV of an aggrading tropical forest becomes more stable around 25 years when their rate of carbon accumulation declines, and that of a tropical peat forest plateaus around 50 years after most of the carbon from peat has been released. The GHGV of tropical forests increases continuously with tE, but is most sensitive to tE at lower tE’s because of strong contributions of GHG release from stored organic matter. Tropical crops and wetland rice decline continually with tE (after an early increase for rice arising from S). Thus, GHGV is particularly sensitive to tE at short time spans and when a large portion of the ecosystems value comes from GHG exchange later in the emissions time frame. Choice of tA has the strongest affect on ecosystems with high GHG exchange later during the time span of interest (Fig. 4b), as the effects of GHG’s released later during the emissions time span take longer to stabilize relative to an initial pulse of CO2. Ecosystems with high

methane flux (FCH4 ; e.g., wetland rice) are especially sensitive to tA, as the global warming potential of CH4 relative to CO2 decreases drastically with tA because of CH4’s short atmospheric lifetime (Forster et al., 2007). For most ecosystems, however, GHGV is not extremely sensitive to tA at analytical time spans greater than 100 years (when tE  50). As with tE and tA, annual discounting has the strongest effect on ecosystems with high GHG flux or slow release of stored organic material (Fig. 4c). The importance of GHG exchange later in the emissions time period declines with increasing annual discount rate such that – at extremely high discount rates – GHGV approaches the value of initial GHG releases from landclearing (i.e., tE 5 0). Thus, decisions regarding the treatment of time are moderately influential for ecosystems whose GHGV’s are attributed mainly to stored organic material (i.e., most native ecosystems with unsaturated soils; Fig. 2) but are extremely influential for ecosystems with high annual GHG flux or slow release of stored organic material (i.e., wetlands, aggrading ecosystems, managed ecosystems). In cases that involve the latter, careful justification of the choice of the emissions time span (tE), the analytical time span (tA), and the choice of a discounting function [w(tA)] are particularly important.

Calculating the full GHG effects of land-use change The full GHG effect of land-use or land-cover change (GHGLUC; Mg CO2-eq ha1) can be readily calculated as the difference in the GHGV’s of the new and old ecosystems: GHGLUC ¼ GHGVnew  GHGVold

ð9Þ

For example, the 50-year cost (with tA 5 100) of burning and clearing a tropical forest with minimal probability of disturbance (GHGV50 100 5 967 Mg CO2-eq ha1) to make way for cropland (GHGV50 100 5121 Mg CO2-eq ha1) would be 1088 Mg CO2-eq ha1, with costs arising both from the lost GHG benefits of maintaining the forest (e.g., C storage) and from the GHG costs of maintaining cropland (e.g., N2O emissions from fertilizer). Conversely, abandoning a tropical cropland such that it becomes an aggrading tropical forest 1 (GHGV50 100 5 529 Mg CO2-eq ha ) would result in a GHG benefit of 650 Mg CO2-eq ha1, with benefits arising both from the cessation of cropping practices and from the CO2 sequestration associated with the growing forest. This approach to calculating GHGLUC captures the full dynamics of land-use or land-cover change (given adequate data) and involves proper weighting for the

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

G R E E N H O U S E G A S VA L U E O F E C O S Y S T E M S timing of emissions. When land use change consists of complete clearing followed by tillage, GHGLUC can be calculated straightforwardly according to Eqn (9). However, for LULCC that does not involve complete clearing followed by tillage (e.g., clearing forest for pasture, conversion of a native forest to a forest plantation, climate change driven land-cover change), it is necessary to adjust the calculation of GHGV of the old ecosystem for the fact that organic matter storage will never be reduced to the baseline level. Specifically, any organic matter pools that will not be disturbed should not be included in the calculations, and decomposition rates should be adjusted to reflect the fact that tillage is not occurring. Legacy effects from the cleared ecosystem are included in its GHGV, and are combined with GHG fluxes from the new ecosystem in a time-specific manner. For example, draining and clearing of peatland to make way for cropland is characterized by an initial release of GHG’s from vegetation followed by decades of enhanced CO2 and N2O release from the organic soil that remains (Kasimir-Klemedtsson et al., 1997), all of which are included in the peatland’s GHGV (see Supporting Information for details). Thus, the GHGV approach to quantifying the GHG effects of LULCC allows accurate quantification of the dynamics of LULCC.

Conclusions

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

Our definition of the GHG value of ecosystems allows straightforward computation of the GHG effect of maintaining ecosystems over a multi-year time frame, expressing their value in terms of storage of organic material that would be released as GHG’s upon land clearing [S; Eqn (5); Fig. 2a], annual uptake or release of GHG’s (F; Fig. 2b), and probable contributions from natural disturbance [Eqns (6)–(8); Fig. 3]. GHGV treats all of these ecosystem-atmosphere GHG exchanges in a time-sensitive manner, thereby providing an appropriate framework for computation of the GHG consequences of any land use decision. It is expressed in units that make it amenable to a wide range of potential applications, and could readily be used to place monetary value on ecosystems simply by multiplying by the market price of carbon ($ Mg1 CO2-eq). GHGV improves upon existing metrics for placing a value on the climate services of ecosystems (Table 1) in that it (1) considers contributions from both flux and storage, (2) uses proper accounting for the timing of emissions over a multiple-year time frame, and (3) allows for the quantification of probable GHG exchange arising from disturbance. Comparison of GHGV with other metrics reveals that properly accounting for storage, flux, disturbance, and the timing of emissions can

Fig. 5 Comparison of GHGV to three commonly used metrics of the GHG contributions of ecosystems (Table 1) for five tropical ecosystem types. Specifically, GHGV50 100 is compared with (1) ‘storage’ – the GHG release that would occur upon clearing of the ecosystem, (2) ‘50 years flux’ – the annual flux multiplied by a 50-year time span of interest, and (3) ‘Storage 1 50 years flux’ – the summation of these two. For these three, CO2-equivalents are calculated based on their global warming potential for a 100-year time horizon (Forster et al., 2007), which corresponds to the tA 5 100 used in computing GHGV50 100).

alter the perceived value of an ecosystem substantially (Fig. 5). Valuation of storage alone underestimates the value of ecosystems that also provide climate services through carbon sequestration (e.g., tropical forest, aggrading tropical forest), but overestimates the value of ecosystems that are GHG sources or would release GHG’s over long time periods upon clearing (e.g., tropical cropland, wetland rice, tropical peat forest). For example, quantifying only the carbon that would be released from clearing a forest (724 Mg CO2-eq ha1) to make way for cropland neglects the costs of displaced carbon sequestration by the forest (252 Mg CO2-eq ha1 50 yr1) and the GHG costs of the cropland (121 Mg CO2-eq ha1 50 yr1), thereby capturing only 66% of the effect of the land use change. More dramatically, assuming that abandoned land that would succeed to forest is of little value because of its low organic matter storage (  4 Mg CO2-eq ha1) neglects the high value arising mainly from its potential for a carbon uptake (529 Mg CO2-eq ha1 50 yr1 for tropical aggrading forests; Fig. 5; see also Righelato & Spracklen, 2007). Valuation of flux alone neglects the value of stored organic matter and is often done in a time-inappropriate manner. While applications counting the cost of landuse/land-cover change typically do not neglect the value of organic matter in high-biomass ecosystems such as forests, they tend to give inadequate attention to the value of organic matter storage in relatively low-biomass ecosystems (e.g., California EPA, 2009; US EPA, 2009). Conversion of moderately grazed tropical pasture to cropland, for instance, would result in

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

11

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

12 K . J . A N D E R S O N - T E I X E I R A & E . H . D E L U C I A

GCB

non-negligible GHG emissions from stored organic matter (177 Mg CO2-eq ha1), which is more influential to its GHGV than is 50 years of annual GHG exchange (48 Mg CO2-eq ha1 50 yr1; Fig. 5). Moreover, counting fluxes in a time-insensitive manner overestimates their impact (Fig. 5). For instance, simply summing the carbon that an aggrading successional forest would sequester over time span of interest ( 618 Mg CO2-eq ha1 50 yr1) overestimates the GHGV of the ecosystem (529 Mg CO2-eq ha1 50 yr1). Arguments that allowing abandoned cropland to regrow to forest or grassland would outweigh the benefits of using the land for biofuels (Righelato & Spracklen, 2007; Pin˜eiro et al., 2009) overestimate the GHG benefits of allowing cropland to revert to natural land. This does not, however, imply that their arguments are fundamentally wrong, as they likewise overestimate the benefits of biofuels. Evaluation of the combined effects of storage and flux over a multi-year time frame miscalculates the value of ecosystems when the timing of emissions is not treated properly (Fig. 5). This arises largely from overestimation of the flux term. For example, a tropical forest would be valued at 1100 Mg CO2-eq ha1 50 yr1 through simple summation of storage and 50 years flux, 1 whereas its GHGV50 100 is only 967 Mg CO2-eq ha . In contrast, if the ecosystem is a GHG source, its value is underestimated through improper treatment of time (e.g., cropland, pasture; Fig. 5). Applications that use this approach – including many biofuel life cycle analyses – effectively place improper weight on flux relative to storage. Many biofuel life cycle analyses (e.g., Fargione et al., 2008; Gibbs et al., 2008; Searchinger et al., 2008; California EPA, 2009; Melillo et al., 2009; Pin˜eiro et al., 2009; US EPA, 2009) may underestimate the GHG costs of biofuels-related land use change relative to the benefits of biofuels production (O’Hare et al., 2009). As a metric with particular strength is calculating the full effects of maintaining an ecosystem over a multipleyear time period, GHGV has many potential applications, and – in many cases – provides a more complete conceptual framework for valuing the GHG services of ecosystems. From a scientific standpoint, GHGV may provide a convenient metric for quantifying the impact of land-use or land-cover changes on the atmosphere. GHGV also may prove useful in forecasting the impacts of land-use policies – such as those relating to biofuels – on future GHG emissions. While assessments of the potential for biofuel production to help mitigate GHG emissions are increasingly recognizing the most important elements of GHGV (e.g., Fargione et al., 2008; Searchinger et al., 2008; California EPA, 2009; Melillo et al., 2009; Smeets et al., 2009; US EPA, 2009), we are not aware of any that include all flux and storage components of GHGV for both old and new ecosystems

involved in both direct and indirect land use changes. Moreover, improper accounting for the timing of emissions may skew their outcomes, as discussed above. The GHGV approach to calculating GHGLUC (Eqn (9)) may provide a more comprehensive and transparent framework for such analyses. Finally, GHGV could also provide a framework for more accurate representation of ecological and atmospheric dynamics in national and international carbon-offsetting programs that reward avoided GHG emissions or GHG sequestration, which require robust treatment of GHG exchanges over time and the possibility of GHG release through disturbance (i.e., ‘volatility’; e.g., Kim et al., 2008). GHGV quantifies the contributions of ecosystems to climate change through ecosystem-atmosphere GHG exchange; however, this does not represent their entire contribution to the climate system, nor does it characterize the full suite of critical ecosystem services. Biogeophysical forcings (e.g., albedo, evapotranspiration, aerosols from biomass burning) affect regional climate and may overshadow the GHG contribution to climate forcing in some biomes (e.g., Betts, 2000; Field et al., 2007; Bonan, 2008). These biogeophysical forcings must be combined with GHGV to quantify the full climate services of ecosystems. Moreover, a suite of ecosystem services that do not directly affect climate – including regulation of water flow and quality, preservation of habitats and biodiversity, production of food, fuel, or fiber, and utilization by native and marginalized peoples (e.g., Foley et al., 2005) – all warrant consideration in the valuation of ecosystems.

Acknowledgements This work was supported by the Energy Biosciences Institute. We thank J. Anderson, S. Davis, J. Endres, C. Smith, and M. Zeri for helpful discussion; an anonymous reviewer for helpful suggestions; and D. Baldocchi, M. Litvak, S. Luyssaert, and T. Meyers for providing data.

References Adler PR, Del Grosso SJ, Parton WJ (2007) Life-cycle assessment of net greenhouse-gas flux for bioenergy cropping systems. Ecological Applications, 17, 675–691. Anderson KJ, Allen AP, Gillooly JF, Brown JH (2006) Temperature-dependence of biomass accumulation rates during secondary succession. Ecology Letters, 9, 673–682. Anderson-Teixeira KJ, Davis SC, Masters MD, DeLucia EH (2009) Changes in soil organic carbon under biofuel crops. GCB Bioenergy, 1, 75–96. Anderson-Teixeira KJ, DeLong JP, Fox AM, Brese DA, Litvak ME (in review). Differential responses of production and respiration to temperature and moisture drive the carbon balance across a climatic gradient in New Mexico. Global Change Biology. Andreae MO, Merlet P (2001) Emission of trace gases and aerosols from biomass burning. Global Biogeochemical Cycles, 15, 955–966. Betts RA (2000) Offset of the potential carbon sink from boreal forestation by decreases in surface albedo. Nature, 408, 187–190. Bonan GB (2008) Forests and climate change: forcings, feedbacks, and the climate benefits of forests. Science, 320, 1444–1449.

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

Q1

G R E E N H O U S E G A S VA L U E O F E C O S Y S T E M S Bond-Lamberty B, Wang C, Gower ST (2004) Net primary production and net ecosystem production of a boreal black spruce wildfire chronosequence. Global

Luyssaert S, Schulze ED, Borner A et al. (2008) Old-growth forests as global carbon sinks. Nature, 455, 213.

Change Biology, 10, 473–487. Bormann FJ, Likens GE (1994) Pattern and Process in a Forested Ecosystem: Disturbance, Development, and the Steady State Based on the Hubbard Brook Ecosystem Study. Springer-Verlag, Berlin. California EPA (2009). Proposed Regulation to Implement the Low Carbon Fuel Standard, Volume I: Staff Report: Initial statement of reasons. Air Resources Board. California Environmental Protection Agency.

Magnani F, Mencuccini M, Borghetti M et al. (2007) The human footprint in the carbon cycle of temperate and boreal forests. Nature, 447, 849–851. Melillo JM, Reilly JM, Kicklighter DW et al. (2009) Indirect emissions from biofuels: how important? Science, 326, 1397–1399. Miles L, Kapos V (2008) Reducing greenhouse gas emissions from deforestation and forest degradation: global land-use implications. Science, 320, 1454–1455. Murty D, Kirschbaum M, McMurtrie RE, McGilvray H (2002) Does conversion of

CCX (2009) Chicago Climate Exchange General Offset Program Provisions. In. Chicago Climate Exchange Inc., Chicago, IL, USA. Chapin F, Woodwell G, Randerson J et al. (2006) Reconciling carbon-cycle concepts, terminology, and methods. Ecosystems, 9, 1041. Davidson E, Ackerman I (1993) Changes in soil carbon inventories following cultivation of previously untilled soils. Biogeochemistry, 20, 161–193.

forest to agricultural land change soil carbon and nitrogen? A review of the literature? Global Change Biol, 8, 105–123. Odum E (1969) The strategy of ecosystem development. Science, 164, 262–270. O’Hare M, Plevin RJ, Martin JI, Jones AD, Kendall A, Hopson E (2009) Proper accounting for time increases crop-based biofuels’ greenhouse gas deficit versus petroleum. Environmental Research Letters, 4, doi:10.1088/1748-9326/4/2/024001.

Dixon RK, Solomon AM, Brown S, Houghton RA, Trexier MC, Wisniewski J (1994) Carbon pools and flux of global forest ecosystems. Science, 263, 185–190. Fargione J, Hill J, Tilman D, Polasky S, Hawthorne P (2008) Land clearing and the biofuel carbon debt. Science, 319, 1235–1238. Fearnside P (2000) Global warming and tropical land-use change: greenhouse gas emissions from biomass burning, decomposition and soils in forest conversion,

Peichl M, Arain MA, Ullah S, Moore TR (2009) Carbon dioxide, methane, and nitrous oxide exchanges in an age-sequence of temperate pine forests. Global Change Biology, 9999. Pin˜eiro G, Jobba´gy EG, Baker J, Murray BC, Jackson RB (2009) Set-asides can be better climate investment than corn ethanol. Ecological Applications, 19, 277–282. Prieme´ A, Christensen S, Dobbie KE, Smith KA (1997) Slow increase in rate of methane

shifting cultivation and secondary vegetation. Climate Change, 46, 115–158. Field CB, Lobell DB, Peters HA, Chiariello NR (2007) Feedbacks of terrestrial ecosystems to climate change. Annual Review of Environment and Resources, 32, 1–29. Foley JA, DeFries R, Asner GP et al. (2005) Global consequences of land use. Science, 309, 570–574. Forster P, Ramaswamy V, Artaxo P et al. (2007) Changes in atmospheric constituents and in radiative forcing. In: Climate Change 2007: The Physical Science Basis.

oxidation in soils with time following land use change from arable agriculture to woodland. Soil Biology and Biochemistry, 29, 1269–1273. Righelato R, Spracklen DV (2007) Carbon mitigation by biofuels or by saving and restoring forests? Science, 317, 902. Robertson GP, Grace PR (2004) Greenhouse gas fluxes in tropical and temperate agriculture: the need for a full-cost accounting of global warming potentials. Environment, Development and Sustainability, 6, 51–63.

Q3

Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (eds Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL), Cambridge University Press, Cambridge, UK and New York, NY, USA. Gibbs HK, Johnston M, Foley JA, Holloway T, Monfreda C, Ramankutty N, Zaks D (2008) Carbon payback times for crop-based biofuel expansion in the tropics: the

Robertson GP, Paul EA, Harwood RR (2000) Greenhouse gases in intensive agriculture: contributions of individual gases to the radiative forcing of the atmosphere. Science, 289, 1922–1925. Roulet NT (2000) Peatlands, carbon storage, greenhouse gases, and the Kyoto protocol: prospects and significance for Canada. Wetlands, 20, 605–615. Scholze M, Knorr W, Arnell NW, Prentice IC (2006) A climate-change risk analysis

Q4

effects of changing yield and technology. Environmental Research Letters, 034001. Guo LB, Gifford RM (2002) Soil carbon stocks and land use change: a meta analysis. Global Change Biology, 8, 345–360. Houghton RA (2007) Balancing the global carbon budget. Annual Review of Earth and Planetary Sciences, 35, 313–347. Houweling S, Ro¨ckmann T, Aben I et al. (2006) Atmospheric constraints on global emissions of methane from plants. Geophysical Research Letters, 33.

for world ecosystems. Proceedings of the National Academy of Sciences, 103, 13116–13120. Searchinger T, Heimlich R, Houghton RA et al. (2008) Use of U.S. croplands for biofuels increases greenhouse gases through emissions from land-use change. Science, 319, 1238–1240. Smeets EMW, Bouwman LF, Stehfest E, Vuuren DPv, Posthuma A (2009) Contribution of N2O to the greenhouse gas balance of first-generation biofuels. Global Change

IPCC (2006) Agriculture, forestry, and other land use. In: 2006 IPCC Guidelines for National Greenhouse Gas Inventories (eds Eggleston S, Buendia L, Miwa K, Ngara T, Tanabe K), Institute for Global Environmental Strategies , Hayama, Japan. ˚ , Klemedtsson L, Berglund K, Martikainen P, Silvola J, Kasimir-Klemedtsson A Oenema O (1997) Greenhouse gas emissions from farmed organic soils: a review. Soil Use and Management, 13, 245–250.

Biology, 15, 1–23. Smith P, Martino D, Cai Z et al. (2007) Agriculture. In: Climate Change 2007: Mitigation Contribution of Working Group III to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change (eds Metz B, Davidson OR, Bosch PR, Dave R, Meyer LA), Cambridge University Press , Cambridge, UK and New York, NY, USA. Tilman D, Hill J, Lehman C (2006) Carbon-negative biofuels from low-input high-

Keppler F, Hamilton JTG, Bra M, Ro¨ckmann T (2006) Methane emissions from terrestrial plants under aerobic conditions. Nature, 439, 187. Kim M-K, McCarl BA, Murray BC (2008) Permanence discounting for land-based carbon sequestration. Ecological Economics, 64, 763–769. Laine J, Laiho R, Minkkinen K, Vasander H (2006) Forestry and boreal peatlands. In: Boreal Peatland Ecosystems (eds Wieder RK, Vitt DH), pp. 331–357. Springer-Verlag, Berlin Heidelberg.

diversity grassland biomass. Science, 314, 1598–1600. Turner MG, Romme WH (1994) Landscape dynamics in crown fire ecosystems. Landscape Ecology, 9, 59–77. UK Parliament (2007). The Renewable Transport Fuel Obligations Order 2007. In: 2007 No. 3072 United Kingdom. UNFCCC (2008). Reducing emissions from deforestation in developing countries: approaches to stimulate action. Report on the Conference of the Parties on its thirteenth session,

Lal R (2004) Soil carbon sequestration to mitigate climate change. Geoderma, 123, 1–22. Law BE, Falge E, Gu L et al. (2002) Environmental controls over carbon dioxide and water vapor exchange of terrestrial vegetation. Agricultural and Forest Meteorology, 113, 97–120. Law BE, Sun OJ, Campbell J, Tuyl SV, Thornton PE (2003) Changes in carbon storage

held in Bali from 3 to 15 December 2007. Part Two: Action taken by the Conference of the Parties at its thirteenth session. United Nations, pp. 8–10. UNFCCC (2009) Copenhagen accord. In: Draft Decision -/CP.15 (ed United Nations Framework Convention on Climate Change). United Nations, Copenhagen, Denmark. US EPA (2009). Regulation of Fuels and Fuel Additives: Changes to Renewable Fuel

and fluxes in a chronosequence of ponderosa pine. Global Change Biology, 9, 510–524. Le Mer J, Roger P (2001) Production, oxidation, emission and consumption of methane by soils: a review. European Journal of Soil Biology, 37, 25. Lenton TM, Held H, Kriegler E, Hall JW, Lucht W, Rahmstorf S, Schellnhuber HJ (2008) Tipping elements in the earth’s climate system. Proceedings of the National Academy of Sciences, 105, 1786–1793. Litvak M, Miller S, Wofsy S, Goulden M (2003) Effect of stand age on whole ecosystem

Standard Program; Proposed Rule. In: (ed. Agency EP). Federal Register, pp. 24904–25133. US House of Representatives (2009). H.R. 2454: American Clean Energy and Security Act of 2009. van der Werf GR, Randerson JT, Giglio L, Gobron N, Dolman AJ (2008) Climate controls on the variability of fires in the tropics and subtropics. Global Biogeochemical Cycles, 22, GB3028.

CO2 exchange in the Canadian boreal forest. Journal of Geophysical Research, 108, D3, doi: 10.1029/2001JD000854. Luyssaert S, Inglima I, Jung M et al. (2007) CO2 balance of boreal, temperate, and tropical forests derived from a global database. Global Change Biology, 13, 2509–2537.

West TO, Marland G (2002) Net carbon flux from agricultural ecosystems: methodology for full carbon cycle analyses. Environmental Pollution, 116, 439. Westerling AL, Hidalgo HG, Cayan DR, Swetnam TW (2006) Warming and earlier spring increase Western U.S. forest wildfire activity. Science, 313, 940–943.

Q2

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

13

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

Q5 Q6

Q7

14 K . J . A N D E R S O N - T E I X E I R A & E . H . D E L U C I A

Supporting Information Additional Supporting Information may be found in the online version of this article: Table S1. Definitions of ecosystem types. Table S2. Organic matter storage (Mg dry matter ha1) in aboveground biomass, root biomass, dead wood, organic layer of litter or peat, soil organic matter (SOM) in the top meter of mineral soil, and SOM vulnerable to loss upon cultivation. Reported are mean values, standard deviations (where applicable), number of sites contributing to the estimate, and references from which the values were obtained. Table S3. Burn characteristics of combustible biomass for various ecosystem types. Listed are the fraction of combustible biomass consumed in a land clearing fire (fcCB; eq. 5), emission factors for CO2, CH4, and N2O (Edx;CB ; eq. 5), categories within original references, and references. Table S4. Decomposition constants for each biomass pool (combustible material, roots, peat, and soil organic matter) across a variety of ecosystem types (eq. 5, 7; also see above). Table S5. Carbon dioxide flux (FCO2 ; kmol CO2 ha1 yr1), or net ecosystem production, in various ecosystem types. Sign convention: negative indicates carbon uptake by ecosystems. Listed are FCO2 estimates, standard deviations (where available), number of sites contributing to the estimate, measurement methodology, and references. Table S6. Methane flux, FCH4 , from soils and enteric fermentation of livestock. Methanogenesis and methanotrophy values (median, n, and category) are as reported in a review by Le Mer & Roger (2001). All units kmol CH4 ha yr1. Table S7. Nitrous oxide flux (FN2 O ; kmol ha1 yr1) for various ecosystem types. Listed are means, standard deviations, number of sites contributing to the estimate, and notes. Table S8. Anthropogenic CO2 emissions associated with agriculture: emissions from fertilizer production, lime application, and fuel use of farm equipment. All are expressed in units of kmol CO2 ha1 yr1. Table S9. Contributions of storage and flux components to GHGV50 100 (50 year emissions time span, 100 years analytical time span) of various ecosystem types. All units Mg CO2-eq/ha. Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.

(BWUK GCB 2220 Webpdf:=04/12/2010 08:42:07 471522 Bytes 14 PAGES n operator=bs.anantha) 4/12/2010 8:42:11 AM

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54

GCB

r 2010 Blackwell Publishing Ltd, Global Change Biology, doi: 10.1111/j.1365-2486.2010.02220.x

2220

Author Query Form _______________________________________________________ Journal Article

GCB 2220

_______________________________________________________ Dear Author, During the copy-editing of your paper, the following queries arose. Please respond to these by marking up your proofs with the necessary changes/additions. Please write your answers clearly on the query sheet if there is insufficient space on the page proofs. If returning the proof by fax do not write too close to the paper's edge. Please remember that illegible mark-ups may delay publication. Query No. Description Author Response

Q1

Author: Please provide the year of publication, volume and page range for reference Anderson-Teixeira et al (in review)

Author: Please provide the editors, chapter title and page range for reference CCX (2009).

Q2 Author: Please provide the volume and page range for reference Gibbs et al (2008).

Q3 Author: Please provide the page range for reference Houweling et al (2006).

Q4 Author: Please provide the page range for reference O'Hare et al (2009).

Q5 Author: Please provide the page range for reference Peichl et al (2009).

Q6 Author: Please provide further details for reference UK Parliament (2007).

Q7

Suggest Documents